In our previous work we proposed a Service-oriented and User-centric Intrusion Detection System (SUIDS) for pervasive/ubiquitous computing. SUIDS meets the special requirements of pervasive computing by adopting a flexible system architecture and user-centric auditing mechanism. In this paper we present the detection methods and experiment results of SUIDS. Specifically, the paper shows how a smoothing constant is used in a user profile to represent the user?s short-term behavior in real-time; and how the deviations between a user?s short-term and long-term behaviors are measured by using a chi-square statistic test. As a result, SUIDS effectively achieved real-time intrusion detection in pervasive computing, with a lightweight, adaptive and heterogeneous system design.